A Technique for Optimal Cloud Traffic Delivery

Abstract

newlineIn demand to growing traffic of cloud and other web services the vendors are offering several newlinealternatives for optimal delivery of the traffic. Emerging technologies like IoT (Internet of newlineThings) and wearable devices like Smart Glass, Smart watch, Smart Bracelet and Smart newlinePlaster produce delay sensitive traffic. Cloud computing services are emerging as supportive newlinetechnologies by providing resources. Most services like IoT require optimizing Cloud Traffic newlinewhich is still an area of research. The most important parameters to optimize cloud traffic are newlinetimely delivery (with low latency), optimized use of bandwidth, optimized routing, and newlineminimization of cost. The current Research focuses on three parameters to optimize cloud newlinetraffic, First Reducing latency of Cloud Traffic Delivery, Better Management of Network newlineEquipment, IP Packet Classification to identify Cloud Traffic. The thesis suggests three newlinesolutions to the above-mentioned problems related to cloud traffic. First is the proposed newlinemodel that includes the establishment of mini data centers in proximity to end users, which newlinesolves the issues related to latency in the delivery of cloud traffic, bandwidth issues, and newlinehence cost issues. Secondly, the research proposes a new network management model in newlineplace of the traditional Simple Network Management Protocol. Thirdly, remarking of IP newline(Internet Protocol) DSCP (Differential Services Code Point) to identify Cloud traffic. newlineSimulation modeling was performed on the first proposed solution through Cloud Analyst newlineand CloudSimSDN. A modeling is performed to Simulate Cloud Mega Data Centres using newlineCloudSim and Cloud MiniData Center using CloudSimSDN. An Algorithm was also newlinedeveloped for packet traversing through CloudSimSDN region. Another Modeling is newlineperformed using CloudAnalyst. The first model is for traditional mega data centers created newlineusing Cloud Analyst, in some regions. The model shows high response times, high VM costs, newlineand high data transfer costs. The second model created through CloudAnalyst includes th